Article
Construction & Building Technology
Sang-Yum Lee, Tri Ho Minh Le, Yeong-Min Kim
Summary: The number of potholes in the world has increased rapidly due to various factors. Predicting and detecting potholes accurately is important for timely maintenance and safety enhancement. This study improves the prediction model by considering independent variables and suggests a computer vision-based system for pothole detection.
DEVELOPMENTS IN THE BUILT ENVIRONMENT
(2023)
Article
Computer Science, Artificial Intelligence
Zelong Kong, Yongquan Chen, Xinping Guan, Xinyi Le
Summary: Object detection is a significant field in computer vision, but the imbalance problem negatively affects performance. This research reveals two sources of imbalance in existing object detection methods and proposes solutions in terms of model architecture and optimization target. By introducing a location scale equilibrium module and a repulsive loss, the proposed method can address the imbalance in the location distribution of objects with different sizes and the representation information of different categories of objects in practical applications.
Article
Chemistry, Multidisciplinary
Sung-Sik Park, Van-Than Tran, Dong-Eun Lee
Summary: Pothole repair is crucial in road maintenance, but current detection methods are labor-intensive and time-consuming. Research shows that YOLOv4-tiny is the best fit model for pothole detection, offering a more efficient and accurate solution for road surface monitoring.
APPLIED SCIENCES-BASEL
(2021)
Article
Construction & Building Technology
Feng Guo, Jian Liu, Chengshun Lv, Huayang Yu
Summary: Currently, there is a need for automatic approaches to detect pavement cracks for maintenance. This study proposes a Transformer-based network for accurate pixel-level pavement crack detection. By utilizing the hierarchical architecture of Swin Transformer and attention module in the UperNet, the proposed model achieves the best performance compared to other models on three public pavement crack datasets. This paves the way for future applications of automatic pavement crack detection using Transformer-based networks.
CONSTRUCTION AND BUILDING MATERIALS
(2023)
Article
Engineering, Multidisciplinary
Caohan Gu, Yi Lu, Mingzhi Chen, Guifang Sun, Zhonghua Ni
Summary: In this study, a reweighting offset bin classification network was proposed to address the issue of inaccurate offset prediction in deep-learning methods used for detecting and locating metal components surface defects. The network discretizes offset values into multiple bins by probability distribution and corrects offset by predicting the probability of these bins, ensuring relatively balanced weight norms for all bins. Comparative experiments demonstrated accurate defect localization with this method.
Article
Computer Science, Artificial Intelligence
Jiajing Liu, Weili Fang, Peter E. D. Love, Timo Hartmann, Hanbin Luo, Lulu Wang
Summary: This research presents an approach that uses computer vision and deep learning to identify and locate multiple unsafe behaviours in digital images from a construction site, achieving good results.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Robotics
Zhihua Xu, Xiaobin Hong, Tianshui Chen, Zhijing Yang, Yukai Shi
Summary: This paper proposes a scale-aware squeeze-and-excitation (SASE) module for lightweight object detection, which addresses the issues of insufficient local feature interactions and multiscale feature fusion. Experimental results demonstrate that the proposed method achieves a good tradeoff between accuracy and model complexity.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Construction & Building Technology
Jinchao Guan, Xu Yang, Ling Ding, Xiaoyun Cheng, Vincent C. S. Lee, Can Jin
Summary: An automated pixel-level pavement distress detection framework integrating stereo vision and deep learning is developed in this study, which establishes multi-feature pavement image datasets based on a multi-view stereo imaging system and proposes a modified U-net deep learning architecture introducing depthwise separable convolution for efficient crack and pothole segmentation. The results show that the 3D pavement image achieves millimeter-level accuracy, and the enhanced 3D crack segmentation model outperforms other models in terms of accuracy and speed, enabling high-precision automated pothole volume measurement.
AUTOMATION IN CONSTRUCTION
(2021)
Article
Robotics
Krzysztof Lis, Sina Honari, Pascal Fua, Mathieu Salzmann
Summary: This study combines road obstacle detection techniques with perspective information to address the issue of diminishing apparent size of obstacles as their distance from the vehicle increases. The results demonstrate that the combination of these two strategies significantly improves obstacle detection performance and outperforms existing methods in terms of instance-level obstacle detection.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Review
Construction & Building Technology
Narges Kheradmandi, Vida Mehranfar
Summary: The article emphasizes the significance of early detection of pavement decay in road maintenance, introduces the application of image processing in this field, and highlights the challenges in pavement image segmentation. The literature review indicates the need for further improvement on existing research results, providing a direction and guidance for future studies.
CONSTRUCTION AND BUILDING MATERIALS
(2022)
Article
Computer Science, Information Systems
Dingwen Zhang, Bo Wang, Gerong Wang, Qiang Zhang, Jiajia Zhang, Jungong Han, Zheng You
Summary: The task of onfocus detection, which aims to identify whether an individual is focused on the camera, is of great importance for criminal investigation, disease discovery, and social behavior analysis. However, due to the lack of large-scale datasets and the challenging nature of the task, research on onfocus detectors is limited.
SCIENCE CHINA-INFORMATION SCIENCES
(2022)
Article
Engineering, Multidisciplinary
Nhat-Duc Hoang, Van-Duc Van-Duc Tran
Summary: This study proposes a computer vision based method for automatic identification of asphalt pavement segregation, and experimental results show that the method has outstanding performance.
Article
Computer Science, Theory & Methods
N. U. Haq, M. M. Fraz, T. S. Hashmi, M. Shahzad
Summary: Automatic detection of weapons is critical for enhancing security, but it is a challenging task due to the variety and occlusion of weapons. In this study, a CNN architecture for Orientation Aware Weapons Detection is proposed, which achieves improved performance compared to existing object detection algorithms. A new dataset with annotated bounding boxes is also provided for further research in this area.
Article
Engineering, Multidisciplinary
Junjie Xing, Minping Jia
Summary: An automatic detection method based on convolutional neural networks is proposed in this paper and its detection performance is evaluated and compared with other models, showing that the method has better performance in real-time automatic detection of workpiece surface defects.
Article
Computer Science, Software Engineering
Mingwen Shao, Wei Zhang, Yunhao Li, Bingbing Fan
Summary: In this paper, a novel label assignment strategy is proposed for object detection, assigning ground-truths to anchor boxes based on their classification scores. Additionally, a branch alignment module is designed to improve classification accuracy and localization precision.
Article
Optics
Hanshen Chen, Yishun Su, Wei He
Summary: The study proposed an enhanced high-resolution crack detection network based on convolutional neural networks, which achieved higher accuracy and robust performance through a series of measures compared to other methods.
Article
Computer Science, Information Systems
Hanshen Chen, Huiping Lin, Minghai Yao